Open SkAI 2025

Dates: September 2–5, 2025

Location: SkAI Hub (172 E. Chestnut St., Suite 3500, Chicago, IL 60611)

We are happy to announce the inaugural conference, “Open SkAI 2025,” of the NSF-Simons AI Institute for the Sky (SkAI Institute), which will take place from Tuesday, September 2 through Friday, September 5, 2025. The Open SkAI Conference is an annual four-day event to enhance and stimulate new Astro-AI research directions. This conference will facilitate new collaborative activities between the AI industry, astronomers, and AI researchers, and stimulate new directions in community software and Astro-AI research. Please feel free to view and share the Conference Flyer. You can also check out our SkAI Navigation Guide.

The week of the conference is between two special weekends in Chicago. The Chicago Jazz Festival will be on the weekend before the conference (Thursday–Sunday, August 28–31, 2025) and is a Labor Day weekend tradition that promotes all forms of jazz through free, high-quality music programming. The festival showcases Chicago’s local talent alongside national and international artists to raise awareness and appreciation for one of the city’s most beloved art forms. The Taste of Chicago will be on the weekend following the conference (Friday–Sunday, September 5–7, 2025) and is a Chicago summertime tradition that showcases the city’s culinary excellence and diversity.

Thank you for the overwhelming interest in Open SkAI 2025! We’re thrilled to share that registration has reached full capacity, and we’re truly grateful for the incredible response from our community.

Click here to view the conference attendee list

Participants agree to follow the SkAI Institute Code of Conduct

Researchers from all levels are encouraged to apply and attend.

Meals: Breakfast, lunch, and coffee breaks will be provided by the conference.

Plenary Speakers:

Federica Bianco, Associate Professor, Department of Physics and Astronomy, University of Delaware

Federica Bianco obtained a bachelor’s degree in Astronomy from the University of Bologna in 2003 and a PhD in Physics from the University of Pennsylvania in 2010; she was a Smithsonian Predoctoral Fellow at the Harvard Smithsonian Center for Astrophysics, a James Arthur postdoctoral fellow at New York University, and a TED fellow. She is an Associate Professor at the University of Delaware (UD) in the Department of Physics and Astronomy and in the Biden School of Public Policy and Administration, and a Resident Faculty member in the Data Science Institute. Federica joined UD in 2019. She is Deputy Project Scientist and Interim Head of Science of the Vera C. Rubin Observatory and was one of the panelists representing the Legacy Survey of Space and Time (LSST) project at the public release of the Rubin First Look images on June 23, 2025, at the National Academy of Sciences in Washington, D.C.

George Karniadakis, Charles Pitts Robinson and John Palmer Barstow Professor of Applied Mathematics and Engineering, Brown University

George Karniadakis received his SM degree (1984) and PhD (1987) from the Massachusetts Institute of Technology (MIT). He was appointed Lecturer in the Department of Mechanical Engineering at MIT in 1987 and, subsequently, he joined the Center for Turbulence Research at Stanford/NASA Ames. He joined Princeton University as Assistant Professor in the Department of Mechanical and Aerospace Engineering and as Associate Faculty in the Program of Applied and Computational Mathematics. George was a Visiting Professor at Caltech (1993) in the Aeronautics Department. He joined Brown University as Associate Professor of Applied Mathematics in the Center for Fluid Mechanics in 1994 and became a full professor in 1996. He has been a Visiting Professor and Senior Lecturer of Ocean/Mechanical Engineering at MIT since 2000. He was Visiting Professor at Peking University (Fall 2007 and 2013). George is a Fellow of the American Association for the Advancement of Science (AAAS, 2019), Society for Industrial and Applied Mathematics (SIAM, 2010–), American Physical Society (APS, 2004–), and American Society of Mechanical Engineers (ASME, 2003–), and an Associate Fellow of the American Institute of Aeronautics and Astronautics (AIAA, 2006–). He received the SIAM/Association for Computing Machinery (ACM) Prize in Computational Science & Engineering (2021), SIAM Ralf Kleinman Award (2015), CFD Award (2007), and inaugural J. Tinsley Oden Medal (2013) by the U.S. Association for Computational Mechanics (USACM). George is the Director of the DOE Center on Physics-Informed Learning Machines (PhILMs) and the lead PI of the Office of the Secretary of Defense/Air Force Office of Scientific Research (OSD/AFOSR) Multidisciplinary University Research Initiative (MURI) of Physics-Informed Neural Networks for partial differential equations (PDEs).

Invited Speakers:

Rachel Mandelbaum, Professor and Interim Department Head, Department of Physics, Carnegie Mellon University

Rachel Mandelbaum is Professor and Interim Department Head in the Department of Physics at Carnegie Mellon University (CMU). Her research interests are predominantly in the areas of observational cosmology and galaxy studies. This work includes the use of weak gravitational lensing and other analysis techniques, with projects that range from development of improved data analysis methods to actual application of such methods to existing data. Currently, Rachel is focusing on data from the Hyper Suprime-Cam (HSC) survey and is working on upcoming surveys, including the Rubin Observatory Legacy Survey of Space and Time (LSST) and the Nancy Grace Roman Space Telescope. She is an active member of the LSST Dark Energy Science Collaboration, for which she served as Spokesperson from 2019–2021, and she is now the lead at CMU of the LSST Interdisciplinary Network for Collaboration and Computing (LINCC) Frameworks Initiative.

Matteo Sesia, Associate Professor, Department of Data Sciences and Operations, University of Southern California

Matteo Sesia is a tenured Associate Professor in the Department of Data Sciences and Operations at the University of Southern California (USC) Marshall School of Business, with a courtesy appointment in Computer Science at the USC Viterbi School of Engineering. His research lies at the intersection of statistics and machine learning, focusing on developing rigorous and practical methods for analyzing high-dimensional and noisy data in settings where traditional modeling assumptions may not hold. A central theme of his work is distribution-free and model-agnostic inference: statistical methods that enable reproducible variable selection and trustworthy uncertainty quantification while working alongside modern machine-learning models. Grounded in applied statistics and built for real-world data, these tools aim to make black-box models and modern AI systems more transparent, reliable, and scientifically defensible. Before joining USC in 2020, Matteo completed his PhD in Statistics at Stanford University, where he was advised by Emmanuel Candès and received the Jerome H. Friedman Applied Statistics Dissertation Award. He is originally from Italy and holds undergraduate and master’s degrees from Politecnico di Torino and Collegio Carlo Alberto.

Talk title: “Conformal Inference for Open-Set and Imbalanced Classification”
This talk presents a novel conformal prediction method for classification in highly imbalanced and open-set settings, where there are many possible classes and not all may be represented in the labeled data. Existing conformal methods require a finite, known label space and typically involve random sample splitting, which implicitly assumes the availability of a sufficient number of observations from each class. Consequently, standard conformal prediction sets may lose coverage if the test set contains previously unseen labels and tend to become inefficient under extreme class imbalance. To retain valid coverage in the presence of unseen labels, we compute and integrate into our predictions a new type of conformal p-values inspired by the classical Good-Turing estimator, which can be used to rigorously test whether a new data point belongs to a previously unseen class. To make more efficient use of imbalanced data, we develop a selective sample-splitting method that partitions training and calibration data based on label frequency. Despite breaking exchangeability, this approach allows maintaining finite-sample coverage through proper reweighting. With both simulated and real data, we demonstrate that our methods lead to prediction sets with valid coverage even in very challenging open-set scenarios with infinite numbers of possible labels and produces more informative predictions under extreme class imbalance.

Fei Sha, Research Scientist, Google Research

Fei Sha’s research interests are Artificial Intelligence/Machine Learning (AI/ML) and AI for Science/Scientific Machine Learning (SciML). At Google Research, he leads a team of scientists and engineers working in those directions. Before joining Google Research, Fei was a Professor of Computer Science and the Zohrab A. Kaprielian Fellow in Engineering at the University of Southern California (USC). He has been recognized with numerous awards and accolades for his innovative work, including being selected as an Alfred P. Sloan Research Fellow in 2013 and receiving an Army Research Office Young Investigator Award in 2012. He has a PhD in Computer and Information Science from the University of Pennsylvania and BSc and MSc degrees from Southeast University (Nanjing, China).

Talk title: “Advances in Probabilistic Generative Modeling for Scientific Machine Learning”
Leveraging large-scale data and systems of computing accelerators, statistical learning has led to significant paradigm shifts in many scientific disciplines. Grand challenges in science have been tackled with exciting synergy between disciplinary science, physics-based simulations via high-performance computing, and powerful learning methods.

In this talk, I will describe several vignettes of our research in the theme of modeling complex dynamical systems characterized by partial differential equations with turbulent solutions. I will also demonstrate how machine-learning technologies, especially advances in generative AI technology, are effectively applied to address the computational and modeling challenges in such systems, exemplified by their successful applications to weather forecast and climate projection. I will also discuss what new challenges and opportunities have been brought into future machine-learning research.

Chenhao Tan, Associate Professor, Department of Computer Science, The University of Chicago

Chenhao Tan is an Associate Professor of Computer Science and Data Science at The University of Chicago, and directs the Chicago Human+AI Lab. He earned his PhD in Computer Science from Cornell University and dual bachelor’s degrees in Computer Science and Economics from Tsinghua University. His research focuses on human-centered AI, communication & intelligence, and AI alignment. His work has been covered by major news media outlets, including The New York Times and The Washington Post. He also won a Sloan Research Fellowship; an NSF CAREER award; an NSF Computer and Information Science and Engineering Research Initiation Initiative (CRII) award; a Google Research Scholar award; research awards from Amazon, IBM, JP Morgan, and Salesforce; a Facebook fellowship; and a Yahoo! Key Scientific Challenges Award.

Talk title: “Structured Creativity in Science: AI for Hypothesis Generation and Research Ideation”
As AI becomes increasingly capable of following instructions and conducting analyses, I believe that scientists will increasingly play the role of selector and evaluator. In this talk, I will share our recent advances in AI-enabled hypothesis generation and research ideation. Rather than treating AI hallucinations as obstacles to eliminate, we leverage data and literature to steer AI creativity toward generating effective hypotheses. I will also introduce HypoBench, a dedicated benchmark for evaluating hypothesis generation, which reveals significant room for potential improvement of current AI models. Finally, I will present ongoing work that formalizes the hypothesis-generation process and research ideation to enable what we call “structured creativity,” a systematic approach to AI-assisted scientific innovation.

The conference is hosted by the U.S. NSF and Simons Foundation-funded SkAI Institute.

The SkAI Institute is one of the National Artificial Intelligence Research Institutes funded by the U.S. National Science Foundation and Simons Foundation. Information on National AI Institutes is available at aiinstitutes.org.